NVENC vs X264 with RTX 2080ti – Streaming & Gaming


One of the biggest hardware improvements
with the new Nvidia RTX graphics card specifically the 2080 and 2080Ti obviously
the 2070 as well, is that they have the new and improved NVENC encoder. That is
the hardware x264 encoder that is built into the GPU and we’ve seen this from a
number of past generation of cards and is something that a lot of people do use
to stream with – but there’s always been the battle of whether that actual output
the video output is good enough to justify using that over X264 with
your CPUs encoder in this video we’re gonna explain the set up this bit and
then we’re gonna talk about the actual hardware performance so you know your
fps in games for you the streamer and then what it’s like for your viewers who
are watching your stream and the local recording if you’re doing any. So let’s
start off with how we’re gonna test everything. First of all we’re testing
three games, CSGO, PUBG and Fortnite, all on their max settings.
We’re testing a 1080p and we’re using OBS to stream to twitch and do a local
recording. In terms of streaming to twitch we’re doing that at 1080p 60 with
a 6000 KB per second bitrate and then we’re also using a separated
encoder or a second instance of the encoder for the indistinguishable file
quality for the local recording. We are also using the very fast preset for the
streaming side of things obviously using the CPU to encode and we’re using as I
said the the same encoder twice so for testing the CPU we’re stress testing
as hard as we can so that is both streaming and including the stream and
encoding a separate local quality file and then we’re also again with NVENC using
NVENC for both the local recording and the stream as well the hardware I’m
using here is an RTX 2080TI and an Intel 9900K – of course this will
give you an idea of how well the 9900K does perform in streaming gaming
but I’m doing a full video comparing the 9900L to the 2700X so if you want to
see that one make sure you subscribe for more videos because that’s coming out
relatively soon And my final bit of explanation for the
video just talking through the performance
results there’s a little bit complicated so let’s run you through a screen shot
of one of the results and talk you through what it all means.
So in the top right hand corner you have the game so in this case it’s Fortnite. Below that
you have the graphics card which is the RTX2080ti, which is the same for all of
them and then below that you have which type of result were looking at. In the
frame here we’ve got average but you can also see 5% mins and 95% max. Now average
is just the average FPS result, the 5% minimums are effectively a
weighted average if you like for a more real-world minimum so well you know
something like fraps will give you a very low minimum number that is the you
know perhaps only happens once or twice in your entire 5-minute benchmarking
session the 5% minimums give you a more reasonable number of what you can expect
your minimums and maximums to be and then a 5% is just the exact same thing
but for the maximums. In terms of the actual bars in the graph itself there
were split into three sections so the leftmost section that one is your
standard FPS results that is the average minimum or maximum FPS results that I
collected in my testing with stocks so no streaming at all X264 which is
obviously using the CPU to encode the stream and your local recording and then
NVENC for the same thing. The next set of results is the FPS difference between
either X264 or NVENC to the stock so that is obviously NVENC compared to the
stock and X264 compared to the stock to give you an idea of how much
performance you lose depending on which one you’re using and then the last
section is the same numbers but in percentages to give you a more relative
idea of how much FPS you’re losing and how much you might actually see a kind
of performance loss and finally there is one mildly anomalous results with using
NVENC actually had higher minimums than the stock this is just generally margin
of error but it will be listed as 0% as a grass can’t show negative percentages but I’ll
include all of the original results on my website so you can take a look at
that and the links down below so all of that explaining done let’s get to
the results starting off with csgo on the averages we can see the x264 is
losing about 43 percent of its total average FPS here with 19 percent only
being lost by NVENC on the minimums it’s pretty much the same story 37 and 16 percent and some pretty healthy minimums of obviously
x264 losing a fair bit and in the maximums it’s a pretty big loss again on
x264 but not so much of a loss on NVENC. In terms of battlegrounds were
looking again a 30% loss for x264 with only 18% loss with NVENC and again some
impressive minimums as well the minimums here are the one erroneous result was
talking about where the minimums are actually higher than stock but again you
can look at the actual numbers themselves and then on the maximums
it’s basically what you’d expect we’re looking a pretty similar results either
way with actually X264 slightly ahead in terms of fortnight
again we’re looking at 24% loss for x264 with only 16% performance
loss with NVENC and again a very similar results with minimums of 25 and
16 and the maximum is holding a pretty similar results who with 17 and 13 so
overall definitely an impressive show for NVENC with it winning basically
everything in comparison. So as you saw NVENC generally beats out x264 in terms of fps
and your kind of your experience in my actual gaming experience the only one
that I really noticed between the two and between stock was x264 using on csgo
that was incredibly choppy as you saw it was a 43% decrease using X264
versus I think the 19% drop the NVENC had in terms of PUBG again all on
averages but you’re looking 18% instead of 30 percent and Fortnite was 16
instead of 24 so generally a much nicer play experience especially as I said on
csgo so now we’ve had a look at what it’s like for you the streamer and your
FPS results and your general gaming experience but was a like for your
viewers what is the end video quality like well I’m using the
locally recorded file because well that’s the easiest one for me to get
access to since it recorded to my own hard drive and that one is obviously
going to be not necessarily the exact same quality as the stream but generally
speaking it will give you a good indication of how well the encoder does
that’s actually processing your videos so I’m gonna be doing a side-by-side
comparison here and I want you guys to comment in the description down below
I’m going to show you footage from all three games and we’re gonna have an a
side and a b-side now the footage won’t be identical because there were multiple
playthroughs and stuff like that but they will be similar in the area on the
map for example for a PUBG and Fortnite and there will be very similar
overall gameplay so you can see which one is which so feel free to let me know
in the comments down below which one you think is X264 and NVENC in the
comments and let you know in a minutes and no cheating which one is which so because I’m editing this after I film
it, obviously, I’m going to leave all the results over there for you take a look
at them and see how well you did in the comments in your comment and let me know
by replying to your own comment how well you did and how many got right there but
for me especially if you didn’t notice anything in particular do let me know as
well I would love to hear that sort of detailed feedback but either way for me
I couldn’t tell the difference I had them side by side on a 1440p monitor and
also this is 1080p footage so in theory it should have been easier for me to
tell because there are stretched out a fair bit but I couldn’t tell the
difference. Now of course that doesn’t mean that you know no one can but from
my experience I can highly recommend nvm see as the encoder of choice for
streamers and gamers who want to record the footage. Of course it’s faster in
terms of fps it was faster in every game that I tested and that’s generally going
to be true because you’re not bottlenecking your CPU in any way and
the x264 encoder that’s built into the cards is one that is a separate part
that is not normally used while gaming so you’re essentially just utilizing an
extra chip effectively rather than trying to make use of the existing chips
that are there and of course when you’re streaming especially if you’re doing any
scaling and stuff like that that will be done by your CPU so while you do still
need a good and you know powerful CPU to stream you don’t necessarily need as you
good of a CPU if you using your graphics card to do of the rendering and I think
it’s generally what I would recommend especially if you’re just starting out.
Now does that mean that I recommend going out and buying an RTX card over
everything else well that’s pretty hard to recommend if you’re already in the
market and you’re already considering buying an RTX card then possibly this
could be a good thing to tip you over the edge
but if you’re someone who isn’t necessarily already looking at this sort
of price point or if you you know you don’t really stream that much and you
just want a decent quality footage then the old gen NVENC still does plenty
fine so I wouldn’t necessarily you know recommend that the huge price
differences that are between the 10 and 20 series right now that
you know this new version of NVENC is worth that but of course that is
definitely up to you and I would love to hear what you think in the comments down
below is this so much improved this is worth an RTX card, or is this more just
a nice feature that if you already have one you’d make yourself let me know in
the comments down below. Of course if this video has convinced you to pick up
on RTX card then I’ll leave some links in the description down below I’ve also
done a full review of these ZOTAC 2080 and 2080ti and
I’ve done a comparison of all three of the cards that are here or at least
they’re the 2080 versions anyway to give you an idea of which one
you might go with if you’re are interested so feel free to take a look
at those videos I’ll leave a card up above or on the end cards down there
otherwise that is kind of it if you want to support this this channel keep me
making these videos that takes me like full days to do the testing just for
this video and I’ve still got more testing for the 2700X to come so feel
free to stick around and make sure to subscribe for those but if you want to
support this work and this channel you can take a look at the links in the
description down below this patreon we can support me directly and thank you to
all of the awesome patrons who already do you can also check out the Amazon and
overclockers UK affiliate links which also massively helped me out and again
thank you to everyone who uses them it genuinely makes a difference in if I can
pay the rent lisman so thank you very much to you guys it’s also a load of
other links this private Internet access which is a good and cheap VPN Humble
Bundle for some cheap games and a lot of other stuff down there so you can take a
look otherwise if you have any questions feel free to leave them in the comments
down below and of course feel free to check out the other videos too well
always thanks for watching hope you enjoyed it and we’ll see you all in the
next video

Leave a Reply

Your email address will not be published. Required fields are marked *